metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: answer
dtype: string
- name: start_position
dtype: int64
- name: end_position
dtype: int64
splits:
- name: test
num_bytes: 10288824
num_examples: 1000
download_size: 446924
dataset_size: 10288824
configs:
- config_name: default
data_files:
- split: test
path: data/test-*